Adenosine 5'-monophosphate (AMP, also known as adenylic acid) is a precursor of adenosine 5'-triphosphate (ATP), the key energy carrying molecule for all living systems. The first committed enzymatic step in the biosynthesis of AMP is the synthesis of adenylosuccinate from inosine 5'-monophosphate (IMP; inosinic acid) and aspartate. The enzyme which catalyzes this step is known as adenylosuccinate synthetase (IMP:L-aspartate ligase(GDP-forming), EC 6.3.4.4, referred to herein as "ADSS").
In E. coli, ADSS is a dimer of identical 48 kD subunits. Its three-dimensional structure has been determined to 2.8 .ANG. resolution (Poland et al., J. Biol. Chem. 268:25334-25342 (1993). In mammalian cells, the ADSS enzyme is present as two isoforms. An acidic form, present in non-muscle tissues, is thought to be involved in de novo production of AMP. A basic form, present in muscle tissue, thought to act as part of the purine nucelotide cycle, which involves interconversion of IMP and AMP with the net result of deaminating aspartate to fumarate (Lehninger, Biochemistry. Worth Publishers, N.Y. (1975), p. 743; Lowenstein, Int. J. Sports Med. 11:S36-S46 (1990).
Genes encoding the ADSS enzyme have been isolated from a variety of species including E. coli (Wolfe and Smith, J. Biol. Chem. 263:19147-19153 (1988)), D. discoideum (Weismuller et al., J. Biol. Chem. 266:2480-2485 (1991 )), mouse (Guicherit et al., J. Biol. Chem. 266:22582-22587 (1991 ); Guicherit et al., J. Biol. Chem. 269:4488-4496 (1994 ), Bacillus subtilus (Maentsaelae and Zalkin, J. Bacteriol. 174:1881-1890 (1992), human (Powell et al., FEBS Lett. 303:4-10 (1992), S. cerevisiae (Genbank accession no. L22185), and Caenorhabditis elegans (EST; Genbank accession no. M75738). However, genes encoding the ADSS enzyme have heretofore not been isolated from any plant species.
Presently, too little is known about the plant ADSS enzyme and its relationship to the ADSS enzymes/genes which have been isolated from other organisms to allow isolation of ADSS encoding genes from any plant species using known approaches.
Methods for isolating genes which are based upon knowledge of the structure of the proteins they encode cannot be applied to plant ADSS genes because too little is presently known about plant ADSS enzymes. Metabolic enzymes such as ADSS are typically very difficult to purify from plants because of their extremely low abundance. In addition, the presence of various phenolic and carbohydrate compounds in plants can interfere with the isolation of pure enzyme with native activity.
In the absence of direct structural information, a number of standard techniques are available for the isolation of proteins and their corresponding genes. Such standard techniques include nucleic acid hybridization and amplification by polymerase chain reaction using oligonucleotide primers corresponding to conserved amino acid sequence motifs. Unfortunately, these techniques would not be expected to be useful for isolation of plant ADSS genes because they rely upon the presence of significant structural similarity (i.e. amino acid and DNA sequence) with known proteins and genes that have the same function. Since there is no significant structural similarity even among the known ADSS genes and proteins from non-plant organisms (see, e.g. Powell et al., FEBS Lett. 303:4-10 (1992)) it is unlikely that these proteins would share any significant structural similarity with plant ADSS proteins.
Another approach that has been used to isolate biosynthetic genes in other metabolic pathways from higher eukaryotes is the complementation of microbial mutants deficient in the activity of interest. For this approach, a library of cDNAs from the higher eukaryote is cloned in a vector that can direct expression of the cDNA in the microbial host. The vector is then transformed or otherwise introduced into the mutant microbe, and colonies are selected that are phenotypically no longer mutant.
This strategy has worked for isolating genes from higher eukaryotes that are involved in several metabolic pathways, including histidine biosynthesis (e.g. U.S. patent application Ser. No. 08/061,644 to Ward et al., incorporated by reference herein in its entirety), lysine biosynthesis (e.g. Frisch et al., Mol. Gen. Genet. 228:287 (1991 )), purine biosynthesis (e.g. Aimi et al., J. Biol. Chem. 265:9011 (1990)), and tryptophan biosynthesis (e.g. Niyogi et al., Plant Cell 5: 1011 (1993)). This strategy has also been used to isolate plant genes including those coding for maize glutamine synthase (Snustad et al, Genetics 120:1111-1114 (1988)), soybean -pyrroline-5-carboxylate reductase (Delauney et al., Mol. Genet. 221:299-305 (1990), maize dihydrodipicolinate synthase (Frisch et al., Mol. Gen. Genet. 228:287-293(1991 )), rape chloroplast 3-isopropylmalate dehydrogenase (Eller et al., Plant Mol. Biol. 18:557-566 (1992); Proc. Natl. Acad. Sci, USA 88:1731-1735 (1991 )), and dihydroorotate dehydrogenase (Minet et al., Plant J. 2:417-422 (1992)).
Microbial mutants thought to be defective in ADSS activity are available (e.g. E. coli purA mutant designated CGCS 5408 and E. coli strains CGCS 4431 and 7039 from E. coli Genetic Stock Center, Yale Univ.; yeast ade12 mutants reported in Doffman, Genetics 6:377-389 (1969)). However, despite the availability of these mutants, application of the complementation technique to isolate cDNAs encoding ADS S enzymatic activity has proven to be unsuccessful for avian (Powell et al., FEBS Lett. 303:4-10 (1992)) and B. subtilis ADSS (Maentsaelae and Zalkin, J. Bacteriol. 174:1881-1890 (1992).
There are several reasons which may explain the failure of this complementation strategy when applied to ADSS, particularly eukaryotic ADSS genes. First, the eukaryotic ADSS cDNA sequence may not be expressed at adequate levels in the mutant microbe, for instance because of codon usage inconsistent with the usage preferences of the microbial host. Second, the primary translation product from the cloned eukaryotic coding sequence may not produce a functional polypeptide, for instance if activity requires a post-translational modification, such as glycosylation, that is not carried out by the microbe. Third, the heterologous protein expressed in E. coli may also be lethal to the cells in which it is expressed, thus rendering its isolation impossible. Fourth, the eukaryotic protein may fail to assume its active conformation in the microbial host, for instance if the protein is normally targeted to a specific organellar membrane system that the microbial host specifically lacks. This last possibility is especially likely for the plant ADSS enzyme, which has been associated in the plant cell with organelles not present in microbial hosts used in the complementation assay (Schubert, Annu. Rev. Plant Physiol. 37:539-574 (1986), and presumably reaches that organellar system as a result of a post-translational targeting mechanism involving both an N-terminal transit sequence, and intrinsic properties of the mature polypeptide (see, e.g. Kohom and Tobin, Plant Cell 1:159 (1989); Li et al., Plant Cell 3:709 (1991); Li et al., J. Biol. Chem. 267:18999 (1992)). Moreover, two other purine biosynthetic genes isolated from plants, 5'-phosphoribosyl-5-aminoimdazole synthetase (Senecoff and Meagher, Plant Physiol. 102:387-399 (1993)) and glycinamide synthetase (Schnorr et al., Plant J. 6:113-121 (1994)) also appear encode proteins that are targeted to the chloroplast.